Method for generating control data for dose adaptation for an application device for applying a crop protection product

Abstract

Disclosed is a computer-implemented method for generating control data for dose adaptation for an application device for applying a crop protection product, the method comprising obtaining, for each of a plurality of portions of an agricultural field, a sensor parameter value of a sensor parameter, wherein the sensor parameter value is derived from the sensor data depicting the respective portion and is indicative of harmful organism characteristics in the respective portion; determining, for each of the plurality of portions, one or more harmful organisms present in the portion; determining a portion-specific dosage for each of the plurality of portions based on the sensor parameter value, on at least one of the one or more harmful organisms, and on the crop protection product; and generating control data for dose adaptation for the application device based on the determined portion-specific dosage and optionally a current location of the application device.

Description

[0001] Method for generating control data for dose adaptation for an application device for applying a crop protection product

[0002] TECHNICAL FIELD

[0003] Disclosed are a method for generating control data for dose adaptation for an application device for applying a crop protection product, a use of the control data for controlling an application device, a system, a computer program product, and a computer-readable medium.

[0004] TECHNICAL BACKGROUND

[0005] Known methods for generating control data for applying a crop protection product will yield control data that cannot account effectively for differences between potential scenarios that may arise in an agricultural field, such as local differences in terms of harmful organisms. Such differences between potential scenarios may result from characteristics of harmful organisms present in the agricultural field being different for different portions of the agricultural field.

[0006] It is therefore an object of the present disclosure to provide a method and corresponding system that generates improved control data for dose adaptation for an application device for applying a crop protection product.

[0007] SUMMARY OF THE INVENTION

[0008] The present disclosure provides a computer-implemented method for generating control data for dose adaptation for an application device for applying a crop protection product, the method comprising: obtaining, for each of a plurality of portions of an agricultural field, a sensor parameter value of a sensor parameter, wherein the sensor parameter value is derived from sensor data depicting the respective portion of the agricultural field and is indicative of harmful organism characteristics in the respective portion; determining, for each of the plurality of portions, one or more harmful organisms present in the portion; determining a portion-specific dosage for each of the plurality of portions based on the sensor parameter value, on at least one of the one or more harmful organisms present in the portion, and on the crop protection product; and generating control data for dose adaptation for the application device based on the determined portion-specific dosage, and optionally on a current location of the application device.

[0009] In particular, the portion-specific dosage is determined based on a predetermined target efficacy, and the method, particularly the determining the portion-specific dosage, comprises selecting a dosage value corresponding to the target efficacy based on an organism-specific dosage-response curve.

[0010] In other words, the present disclosure provides a computer-implemented method for generating control data for dose adaptation for an application device for applying a crop protection product, the method comprising, for a portion of an agricultural field and / or for a site of an agricultural field: obtaining a respective sensor parameter value of a sensor parameter, wherein the respective sensor parameter value is derived from sensor data depicting the respective portion and is indicative of harmful organism characteristics in the respective portion of the agricultural field; determining one or more harmful organisms present in the respective portion of the agricultural field; determining a portion-specific dosage for the respective portion based on the respective sensor parameter value, on at least one of the one or more determined harmful organisms present in the respective portion, and on the crop protection product; and generating control data for dose adaptation for the application device based on the determined portion-specific dosage, and optionally on a current location of the application device.

[0011] In particular, the portion-specific dosage is determined based on a predetermined target efficacy, and the method, particularly the determining the portion-specific dosage, comprises selecting a dosage value corresponding to the target efficacy based on an organism-specific dosage-response curve.

[0012] In yet other words, the present disclosure provides a computer-implemented method for generating control data for dose adaptation for an application device for applying a crop protection product, the method comprising, for a plurality of portions of an agricultural field: obtaining a portion-specific sensor parameter value of a sensor parameter, wherein the portion-specific sensor parameter value is derived from sensor data depicting the respective portion and is indicative of harmful organism characteristics in the respective portion; determining one or more harmful organisms present in the respective portion; determining a portion-specific dosage for the respective portion based on the portion-specific sensor parameter value, on at least one of the one or more harmful organisms present in the respective portion, and on the crop protection product; and generating control data for dose adaptation for the application device based on the determined portion-specific dosage, and optionally on a current location of the application device.

[0013] In particular, the portion-specific dosage is determined based on a predetermined target efficacy, and the method, particularly the determining the portion-specific dosage, comprises selecting a dosage value corresponding to the target efficacy based on an organism-specific dosage-response curve.

[0014] In particular, the current location of the application device and the respective relative position of the spray devices on the application device may be taken into account when generating the control data. For example, the control data for dose adaptation may be control data generated based on portionspecific dosage and a current location of the application device and relative position of the spray devices on the application device. As an example, the control data may allow for individually controlling the spray devices of the application device dependent on their location relative to the respective portion(s) and the respective portion-specific dosage.

[0015] The present disclosure further provides the use of the control data generated by the method of the present disclosure for controlling an application device to apply crop protection product in accordance with the control data. The present disclosure further provides a system comprising a computing system configured to carry out the method of the present disclosure.

[0016] In other words, the present disclosure provides a system comprising a computer system, the computer system configured to generate control data for dose adaptation for an application device for applying a crop protection product, wherein generating control data comprises: obtaining, for each of a plurality of portions of an agricultural field, a sensor parameter value of a sensor parameter, wherein the sensor parameter value is derived from sensor data depicting the respective portion of the agricultural field and is indicative of harmful organism characteristics in the respective portion; determining, for each of the plurality of portions, one or more harmful organisms present in the portion; determining a portion-specific dosage for each of the plurality of portions of the agricultural field based on the sensor parameter value, on at least one of the one or more harmful organisms present in the portion, and on the crop protection product; and generating control data for dose adaptation for the application device based on the determined portion-specific dosage, and optionally a current location of the application device.

[0017] In particular, the portion-specific dosage is determined based on a predetermined target efficacy, and the method, particularly the determining the portion-specific dosage, comprises selecting a dosage value corresponding to the target efficacy based on an organism-specific dosage-response curve.

[0018] In particular, the computing system may generate the control data by executing a first module, a second module, a third module, and a fourth module, wherein the first module is configured to obtain the sensor parameter value, wherein the second module is configured to determine the one or more harmful organisms present in the portion, wherein the third module is configured to determine the portion-specific dosage, and wherein the fourth module is configured to generate the control data for dose adaptation.

[0019] In particular, the portion-specific dosage is determined based on a predetermined target efficacy, and the method, particularly the determining the portion-specific dosage, comprises selecting a dosage value corresponding to the target efficacy based on an organism-specific dosage-response curve.

[0020] The computer system may be configured as a distributed computer system. In particular, the method may be executed by different computing devices of the distributed computer system. In particular, the distributed computer system may comprise one or more servers and one or more client devices and the method may be executed in part by at least one of the servers and in part by at least one of the client devices. At least one of the client devices may be part of or arranged at the application device.

[0021] The system may comprise one or more devices other than the computer system. The system may comprise one or more, in particular all, components of an application device, particularly one or more spray nozzles. The system may comprise one or more sensors configured to acquire the sensor data. The present disclosure further provides a computer program product comprising instructions which, when carried out by a computer, cause the computer to carry out and / or control the method of the present disclosure.

[0022] The present disclosure further provides a computer-readable medium having stored thereon instructions which, when carried out by a computer, cause the computer to carry out and / or control the method of the present disclosure.

[0023] Details on the computer-implemented method for generating control data for dose adaptation for an application device for applying a crop protection product according to the present disclosure will be described below.

[0024] The determining the one or more harmful organisms present in the portion, in the present disclosure, may be a determining which type or types of harmful organisms are present in the portion.

[0025] The term “agricultural field”, according to the present disclosure, is to be understood broadly and may refer to a piece of land planted with a crop. Some non-limiting examples for crops include corn, sunflower, soy, or the like.

[0026] The agricultural field may be divided into a set of portions of the agricultural field, wherein the set of portions encompasses the entire agricultural field. The plurality of portions, according to the present disclosure, may comprise all portions of the set of portions or may comprise only a subset of the set of portions. In other words, the method for generating control data may be carried out for the entire agricultural field or only for part of the agricultural field.

[0027] The computer-implemented method for generating control data for dose adaptation according to the present disclosure may comprise individually obtaining, for the respective portion, the respective sensor parameter value, individually determining, for the respective portion, the one or more harmful organisms present, and individually determining, for the respective portion, the respective portionspecific dosage.

[0028] In an example the plurality of portions, according to the present disclosure, may all have substantially the same size. Alternatively, the plurality of portions may comprise portions of different sizes.

[0029] In general, in addition to crop, harmful organisms will be present on an agricultural field. Harmful organisms, according to the present disclosure, may comprise at least one of weeds, fungi, and pests. Pests may comprise insects.

[0030] In the present disclosure, weeds, fungi, and pests are each considered to be a category of harmful organism. Within each category of harmful organism, there may be sub-categories of harmful organisms. These sub-categories will be referred to as types of harmful organisms hereinbelow.

[0031] A subcategory might be panicum like plants, a non-taxnomic group containing genera, inter alia: Setaria, Panicum or Echinochloa. The plant taxonomy with their sub-categories is exemplary depicted below.

[0032] Crop protection products may be applied to an agricultural field.

[0033] A crop protection product may be a product configured to prevent and / or mitigate growth of harmful organisms. A crop protection product may comprise at least one of a fungicide, herbicide, and pesticide.

[0034] More specifically, crop protection products may comprise chemical products such as fungicides (e.g. of the chemical classes of the triazoles, the carboxamides, methoxyacrylates etc.), plant growth regulators (e.g. of the chemical classes of the triazoles, gibberilin inhibitors, quartemay ammonium salts etc.), insecticides (e.g. of the chemical class of the pyrethroides, juvenile hormone analoges etc.), herbicides, acaricides, molluscicides, nematicides, avicides, rodenticides, repellants, bactericides, biocides, safeners, water or any combination thereof. Further biological products such as microorganisms useful as fungicide (e.g. Bacillus subtilis, Bacillus amyloliquefacients), herbicide (bioherbicide), insecticide (bioinsecticide e.g. Bacillus thuringiensis), or plant growth promoting bacteria (e.g. Bacillus amyloliquefacients, Bacillus pumilus etc.) acaricide (bioacaricide), molluscicide (biomolluscicide), nematicide (bionematicide- e.g. Bacillus firmus), avicide, piscicide, rodenticide, repellant, bactericide, biocide, safener, total herbicide / burndown herbicide (e.g. pelargonic acid), defoliant, desiccant, or any combination thereof.

[0035] A crop protection product of the present disclosure may be a single type of crop protection product or may be a combination of different types of crop protection products.

[0036] The crop protection product of the present disclosure is also referred to as “tank mix” hereinbelow. The crop protection product or tank mix may be characterized by the type or types of crop protection product and optionally by the ratio of types of crop protection products. In other words, the protection product may be generated by mixing a plurality of different types and / or of different components of protection products. The different types of protection products may be stored in different tanks and / or containers and the crop protection product may be mixed substantially at the moment of applying the crop protection product to the field, in particular to the portion of the field. Such a mixed crop protection product may be referred to as “tank mix”. In an example, a product A is to be applied to the portion at a dosage of 2 l / ha and a product B is to be applied to the portion at a dosage of 4 l / ha. In this example the mixing ratio of product A / product B is 1 / 2.

[0037] The determining the portion-specific dosage based on the crop protection product may, for example, comprise determining the portion-specific dosage based on one or more of the crop protection product type or types, components of crop protection products, ratios of types or components of crop protection products.

[0038] As outlined above, the sensor parameter is indicative of harmful organism characteristics. In other words, the sensor parameter allows for parametrizing harmful organism characteristics. In this way a characteristic of a harmful organism may be mapped to sensor parameter. The sensor parameter may represent the characteristics of a harmful organism in a form that makes it possible to process the parameter by an electronic processing device. The harmful organism characteristics for the agricultural field can be described in a spatially resolved manner, particularly if the location of parameterizing is registered, by determining the values of the sensor parameter for respective portions of the agricultural field, in particular by geo coordinates of the respective portions. Harmful organism characteristics may comprise at least one of morphological parameters of the harmful organism like height and / or width and / or leaf form and / or plant form and / or color characteristics, growth stage of the harmful organism, biomass of the harmful organism, coverage area of the harmful organism, and / or abundance of the harmful organism.

[0039] Accordingly, the sensor parameters may, for example, be indicative of at least one of morphological parameters of the harmful organism like height and / or width and / or leaf form and / or plant form and / or color characteristics, growth stage of the harmful organism, biomass of the harmful organism, coverage area of the harmful organism. For example, if an image sensor and / or a camera is used the parameters are derived from the harmful organism characteristics by way of image processing. In a more detailed example, the height, width and / or leaf form and / or plant form of a harmful organism, in particular of weed, may be characterized by a point cloud and may be derived from lidar sensor and / or imaging sensor data Further it may also be an output of one or multiple model(s). A point cloud depicts the relation of data points in a 3 dimensional space. Further, geometric principles such as the Pythagoras formula may be used to calculate the distance between two object points.

[0040] Sensor parameter values may comprise values comprised in sensor data generated by the sensor. Alternatively or in addition, sensor parameters may be derived sensor parameters. Values of the derived sensor parameters, referred to as derived sensor parameter values, may be obtained by providing sensor data and / or pre-processed sensor data to a model. In other words, sensor parameter values may be derived from sensor data and / or pre-processed sensor data using a model. The derived sensor parameter value may be a second derived sensor parameter value and may be obtained by inputting a first derived sensor parameter value, which may also be referred to as intermediate sensor parameter value, into a model. The intermediate sensor parameter value may, for example, be derived from the sensor data using a first model. The above pre-processed sensor data may comprise intermediate parameter values.

[0041] The derived sensor parameter, particularly the second derived sensor parameter, may be a growth stage, biomass and / or coverage. If the growth stage parameter is a second derived sensor parameter, the first derived sensor parameter may comprise a height, width, leaf / plant form and / or biomass. For example, a growth stage parameter of the harmful organism may be determined by a growth model which takes one or more other sensor parameters as input, like height, width, leaf / plant form and / or biomass. A biomass parameter may be derived by height, width, leaf / plant form and / or growth stage as input to a biomass model. A coverage parameter may be determined by the covered area of the identified harmful organism. The covered area may be determined via optical parameters like green pixels, index values or wavelength values in combination with image processing methods like an automatic threshold method, e.g. the OTSU threshold method, and / or models. Index values are in an example dimensionless numbers of wavelengths calculated with each other.

[0042] The value of the sensor parameter for a portion of the agricultural field is referred to as sensor parameter value. The sensor parameter value is a measure that reflects properties and / or characteristics of the harmful organism for the respective portion. The sensor parameter value may differ between different portions and / or at different locations dependent on the appearance of the respective harmful organism at the respective location.

[0043] The term sensor parameter value is to be understood broadly and may entail data of any type, may originate from any sensor type or sensor model, and may be sensor data output directly from a sensor or may result from processing the sensor data output by the sensor. Accordingly, the sensor parameter may have any unit or unit-less. As an example, the sensor parameter may have a specific value that may be expressed by a number or an alphanumerical identifier, such as a textual identifier, e.g., “zero”, “low”, “medium”, “high”.

[0044] As outlined above, the sensor parameter value is derived from sensor data depicting the respective portion of the agricultural field.

[0045] Sensor data comprises data obtained by a sensor. In particular, sensor data may comprise sensed values obtained by the sensor.

[0046] As an example, the sensor may comprise an optical sensor, such as a camera, a infrared detector, radar (radio detection and ranging) device, a lidar (radio detection and ranging) device and / or a scanner. The sensor may, in particular, be a sensor sensing any part of the optical spectrum, optionally including visible and / or infrared and / or ultraviolet. The sensor data may comprise optical image data, such as camera image data or scanner image data that may be converted into parameter values. Other sensors and respective sensor data are conceivable.

[0047] The sensor parameter values being derived from the sensor data may comprise extracting the sensor parameter values comprised in the sensed values obtained by the sensor or calculating the sensor parameter values from at least some of the sensed values obtained by the sensor.

[0048] Sensor data depicting a portion of the agricultural field may comprise data obtained by a sensor for said portion of the agricultural field.

[0049] Different portions of the agricultural field may have the same or different sensor parameter values. In other words, for different portions of the agricultural field, the sensor parameter values may be different.

[0050] Sensor parameters can be defined in accordance with the scenario at hand. For example, a sensor parameter may be defined that represents or parametrizes a specific harmful organism characteristic selected in accordance with the scenario at hand. As an example, harmful organism coverage may be selected, and a sensor parameter may be defined that represents the harmful organism coverage. For example, such a sensor parameter value may be one of “zero”, “low”, “medium”, and “high” depending on the extent of coverage derived from the sensor data. As an example, senor data, e.g. images, may be processed. As an example, an image of a portion of a field may be processed to determine a value quantifying the harmful organism coverage for the portion as depicted in the image. Then a categorization step may be carried out, where different categories correspond to different ranges of values quantifying harmful organism coverage. The categorization may comprise determining, for the value quantifying the harmful organism coverage as depicted in the image, in which range this value is found and categorizing the image in the corresponding category. Each category may correspond to a sensor parameter value. For example, there may be four categories associated with the values “zero”, “low”, “medium”, and “high”.

[0051] As seen above, a sensor parameter can be defined so as to represent or parametrize a specific harmful organism characteristic. Thus, the sensor parameter value of the sensor parameter will be indicative of the specific harmful organism characteristic. The characteristic, as explained above, may be associated with coverage, and / or abundance, and / or growth state of the harmful organism(s). As already explained above, the sensor data may comprise or correspond to parameter values, or the sensor parameter values may be calculated from the sensor data.

[0052] Examples for harmful organisms are weed like Alopecurus myosuroides (ALOMY), Amaranthus palmeri (AMAPA), Lolium multiflorum (LOLMU) or Echinochloa crus-galli (ECHCG).

[0053] Determining harmful organism characteristic(s) from image data may be carried out using known image processing techniques, which may, for example, comprise using machine learning models trained on labelled image data, and / or may comprise using known morphological properties and identifying them in images. Certain harmful organisms may, for example, be reflected in the images with characteristic shapes and / or colors and / or distributions.

[0054] As will be understood from the above explanations, the sensor parameter values obtained for the different portions of the agricultural field are indicative of the characteristics of the harmful organisms in said portions. Consequently, the sensor parameter values are suitable for use in determining a dosage of a specific crop protection product to be applied to the respective portions of the agricultural field.

[0055] The method of the present disclosure may make use of this allocation of a parameter value to a specific location and derives a corresponding dosage. Specifically, the method of the present disclosure comprises determining a portion-specific dosage of the crop protection product based, among others, on the sensor parameter value. In particular, the method may base determining the portion-specific dosage on only a single sensor parameter value per portion.

[0056] Based on the desired efficacy and based on the available tank mix, a dosage for the tank mix, i.e., the crop protection product, may be determined. This dosage may be converted into control data that are adapted to control an application or treatment device, e.g. a spraying device like a tractor or drone.

[0057] The portion-specific dosage may be determined individually for each of the portions. The portionspecific dosage may be the same or different for different portions. The portion-specific dosage of the crop protection product may, for example, be expressed by a volume per area, e.g. liter per surface area or l / ha. The dosage may be set by selecting a threshold value and / or by setting up a duty cycle for a PWM (pulse width modulation) device.

[0058] The determining of the portion-specific dosage may comprise providing input data to a method for determining a portion specific dosage, the input data comprising the sensor parameter value for the respective portion, data indicating the crop protection product, and data indicating the harmful organism present in the respective portion. The portion-specific dosage may be comprised in output data of the method that determines the portion-specific dosage.

[0059] The data indicating the crop protection product and / or tank mix may be obtained via a user input specifying the crop protection product. Alternatively, the data indicating the crop protection product may be derived from a user input selecting an operation mode in accordance with the selected operation mode or derived from a computer-readable label associated with the crop protection product.

[0060] The data indicating the harmful organism present in the respective portion may be obtained via a user input specifying one or more harmful organisms present in the respective portion and / or plurality of portions (group of portions). In particular, the user input may be provided at field level or at a level relating to a group of portions of the field. This allows for avoiding that the user has to input many individual inputs for individual portions.

[0061] In other words, the user may assess actual presence and type(s) of harmful organism(s) on the field and provide corresponding user input, e.g. via a user interface.

[0062] A user may have the option to select one or more types of harmful organism. For example, a user may select among a list of potential harmful organism types one or more harmful organisms considered to be present in the portion or group of portions, particularly entire field.

[0063] For example, the type of harmful organism that requires the highest dosage may be determined automatically, e.g. by a computing system, as the harmful organism present in the portion or group of portions, particularly entire field. The determined type of harmful organism may then be used for determining the portion-specific dosage.

[0064] Alternatively or in addition to a manual user input, data indicating the harmful organism present in the respective portion may be obtained by processing sensor data, particularly the sensor data used for deriving the sensor parameter value. In an example, both user input and / or sensor parameter values are stored in a register and / or a database and are provided as input to the method via this register and / or database.

[0065] Alternatively or in addition to the sensor data used for deriving the sensor parameter value, data indicating the harmful organism present in the respective portion may be obtained by processing additional sensor data. The additional sensor data may be obtained by the same or a different sensor than the sensor used for obtaining the sensor data used for deriving the sensor parameter value. The additional sensor data may be data obtained prior to or at the same time as the sensor data used for deriving the sensor parameter value.

[0066] For example, as already described above, sensor data may allow for deriving one or more types of harmful organism(s) actually present in a portion or a group of portions, in particularly the entire field. Similarly to the user input, in case a plurality of types of harmful organisms are actually present, a single harmful organisms may be selected for use in the method, e.g. the one with the highest dosage demand.

[0067] As will be understood from the above, whereas the harmful organism characteristics are derived from sensor data, this is optional for the step of determining the presence of harmful organisms. This may have different reasons. For example, a user may already be aware of harmful organisms present in the different portions of the agricultural field, for example based on previous measurements or user experience. Alternatively or in addition, the presence of harmful organisms may also be modelled.

[0068] The harmful organism characteristics, however, are derived from sensor data. Making this distinction can be advantageous. For example, the characteristics of a harmful organism may be more likely to change over time than the mere presence of the harmful organism. It may also, in some cases, be less complex to determine the characteristics, such as coverage, from sensor data, particularly in real-time, than it is to identify the harmful organism.

[0069] From the sensor data, the quantity, size and / or coverage may be determined.

[0070] Optionally, in addition, the type of organism may be detected. In other words, if the type of harmful organism is detected, it may be possible to identify the crop protection product and / or the tank mix that is able to treat in particular this specific identified type, e.g. by using the corresponding mode of action. However, it is not necessary to use detection, as explained above, and user input may instead be used in case a harmful organism type is known to a user or assumed to be present for other reasons.

[0071] A detection may be a two-step approach. In a first step the type is identified and in a second step the appropriate dosage is determined based on the provided type. The type of harmful organism may be determined by sensor and / or by manual user input. Using sensors may allow for determining the harmful organism type per portion whereas the manual input may only allow for a field level and / or group of portions level.

[0072] As explained above, the method of the present disclosure comprises generating control data for dose adaptation for the application device based on the determined portion-specific dosage, and optionally a current location of the application device. For example, the control data for dose adaptation may be control data generated based on portionspecific dosage and a current location of the application device and relative position of the spray devices on the application device. Portion-specific application may comprise individually controlling the spray devices of the application device dependent on their location. Multiple portions may be present per application device. Thus the portion specific dosage may be applied by one or multiple spray devices.

[0073] An application device, according to the present disclosure, may be any device configured to apply crop protection product to an agricultural field. Such devices may comprise spray nozzles connected to a crop protection product tank mounted on a vehicle.

[0074] The application device may be configured to move through the agricultural field. The application device may be a ground-borne vehicle or an air-borne vehicle, e.g. a rail vehicle, a robot, an aircraft, an unmanned aerial vehicle (UAV), a drone, or the like. The application device can be an autonomous or a non-autonomous application device.

[0075] Application devices or treatment devices may be configured to apply crop protection product(s) in accordance with control data. Control data may be data configured for use in controlling operation of the application device. Control data for dose adaptation may be seen as a specific type of control data. Control data for dose adaptation may, for example, be seen as data configured for use in controlling operation of the application device such that the dose applied by the application device is adapted. Control data need not have any specific data format. How the dose may be adapted will be explained in more detail below. Control data may be provided as a control signal and / or as an application map.

[0076] The control data for dose adaptation may be configured such that the dose is adapted in such a manner that, for a respective portion of the agricultural field, the portion-specific dosage is applied.

[0077] As an example, in order to determine, during application of the crop protection product, the portion of the agricultural field and the corresponding portion-specific dosage, the current location of the application device may be taken into account. That is, for example, input data specifying the current location of the application device may be used to determine the current portion or portions and / or the upcoming portion or portions of the agricultural field. Control data may be generated that that adapts dose in accordance with the portion-specific dosage of the current portion or portions and / or the upcoming portion or portions. Input data specifying the current location may, for example, comprise data received from a GPS sensor. The current and / or upcoming portion(s) may relate to the portion(s) are those in reach of the spray devices of the application device.

[0078] Thus, to summarize, the present disclosure provides a method that allows for dose adaptation of an application device that takes into account harmful organisms that are present and the crop protection product and also harmful organism characteristics, represented by a sensor parameter value for individual portions.

[0079] Accordingly, various scenarios in the agricultural field in terms of harmful organism characteristics can be effectively accounted for.

[0080] Specifically, the method of the present disclosure allows for providing improved control data for dose adaptation for an application device for applying a crop protection product.

[0081] Particularly, it may be sufficient to use only a single sensor parameter value in addition to the crop protection product and at least one of the one or more harmful organisms present in the portion in order to obtain improved control data for dose adaptation for an application device for applying a crop protection product. In particular, according to the present disclosure, for a respective portion, the portion-specific dosage may be determined based on a single sensor parameter value, on a single harmful organism, and on the crop protection product.

[0082] According to the present disclosure, the portion-specific dosage may be determined based on a predetermined target efficacy. The target efficacy may be determined based on the sensor parameter value and the crop protection product(s), e.g., tankmix.

[0083] The target efficacy may be a sensor-parameter-value-specific target efficacy. That is, for each sensor parameter value there may be a respective target efficacy associated with the sensor parameter value. The target efficacy associated with the sensor parameter value is referred to sensor-parameter-value-specific target efficacy. Nonetheless, the target efficacy may also be based on the crop protection product(s) as explained above.

[0084] The method may comprise selecting a dosage corresponding to the target efficacy based on an organism-specific dosage-response curve. The dosage response curve may represent the efficacy of the crop protection product for a given type of harmful organism, a growth stage, and / or abundance of the harmful organism, and / or biomass of the harmful organism. The curve may not have a specific growth stage, abundance or biomass and cover a broad range of characteristics.

[0085] In other words, determining the portion-specific dosage may comprise using a dosage-response curve as an input and using a predetermined target efficacy as an input. For example, the dosageresponse curve may be seen as defining the functional relation between the dosage of the crop protection product and the efficacy of the crop protection product for a given type of harmful organism. Accordingly, based on the dosage-response curve, a dosage that corresponds to the predetermined target efficacy can be determined.

[0086] An advantage of using a target efficacy and dosage-response curve is an improved accuracy of determining a dosage.

[0087] Depending on the application case at hand, efficacy may be expressed in different measures. For example, efficacy may be stated as a gain in agricultural yield; that is, the bigger the yield, the greater the efficacy. It may also be expressed as a reduction in seed production, weed tillering, leaf area reduction or leverage chemical content analysis, and the delta of the corresponding tracer. Efficacy may be as well stated as a percentage of biomass decrease.

[0088] The predetermined target efficacy may be retrieved from data associating target efficacies with sensor parameter values. Said data may be retrieved from a data storage. For example, a plurality of sensor parameter values and their associated target efficacies may be retrieved from a data storage. Alternatively, a functional relationship between the sensor parameter and the target efficacy may be retrieved from a data storage. This similarly applies for the functional relationship between target efficacy and crop protection product.

[0089] The dosage-response curve may be retrieved from a data storage, for example in response to a user selection of the dosage-response curve or by automatically retrieving a dosage-response curve that is associated with the crop protection product.

[0090] Optionally, determining the portion-specific dosage may be carried out so as to comply with a maximum allowed dosage. The maximum allowed dosage may be retrieved from a data storage or obtained by a user input or via a computer-readable label. The maximum allowed dosage may, for example, be set by a regulation and / or by a supplier. Accordingly, where a portion-specific dosage is determined based on the predetermined target efficacy, after retrieving the dosage associated with the target efficacy, a step of checking whether the maximum allowed dosage is exceeded by the retrieved dosage may be carried out. In case the maximum allowed dosage is not exceeded, the retrieved dosage may be used as the portion-specific dosage and in case the maximum allowed dosage is exceeded, the maximum allowed dosage may be used as the portion-specific dosage.

[0091] Alternatively or in addition, determining the portion-specific dosage may be carried out so as to comply with a minimum allowed dosage. The minimum allowed dosage may be retrieved from a data storage or obtained by a user input or via a computer-readable label.

[0092] According to the present disclosure, the method may comprise one or more determination modes for determining the portion-specific dosage comprising a first determination mode and / or a second determination mode. The determination mode may, for example, be determined based on a user input selecting the determination mode.

[0093] Determining the portion-specific dosage using the first determination mode may comprise: retrieving a predetermined target efficacy associated with the sensor parameter value, for example as described above, retrieving an organism-specific dosage-response curve, which may represent the efficacy of the crop protection product for a given type of harmful organism, for example as described above, determining the dosage value associated with the target efficacy from the dosage-response curve, for example by selecting a dosage value corresponding to the target efficacy based on the dosage-response curve as described above, and selecting said dosage value as portion-specific dosage.

[0094] Features, examples, and advantages associated using a dosage-response curve have been outlined in detail above.

[0095] Optionally, selecting the dosage value as portion-specific dosage may depend on a maximum allowed dosage not being exceeded, as explained above.

[0096] Determining the portion-specific dosage using the second determination mode may comprise: retrieving a predetermined relative dosage value, such as a percentage, associated with the sensor parameter value or absolute dosage value associated with the sensor parameter value, and if a relative dosage value is retrieved, the second determination mode may further comprise determining the portion-specific dosage by multiplying the relative dosage value with a baseline dosage value.

[0097] For example, each sensor parameter value may be associated with a predetermined relative dosage value. For example, value pairs of sensor parameter and relative dosage may be stored in a data storage. A relative dosage value may be a dosage value related to a baseline dosage value, for example. The baseline dosage is to be understood to be a non-zero value.

[0098] The baseline dosage value may be a predetermined baseline dosage value stored in a data storage. In order to determine the portion-specific dosage, the baseline dosage value may be retrieved, e.g. from a data storage.

[0099] The relative dosage value and the baseline dosage value, in principle, may be defined in any suitable manner. As an example, the baseline dosage value may correspond to a maximum dosage. The relative dosage values may be predetermined percentages of the baseline dosage value, each at most 100 percent. Alternatively, the baseline dosage may correspond to a minimum dosage and the relative dosages may be predetermined multiples of the baseline dosage value. An advantage of the second operation mode is that it does not require that a dosage-response curve exists for the crop protection product and harmful organism. It is sufficient to know a limited number of mutually associated values of predetermined relative dosage and sensor parameter. These mutually associated values may also be referred to as value pairs of a predetermined relative dosage value and an associated sensor parameter value.

[0100] Optionally, selecting the dosage value as portion-specific dosage may depend on a maximum allowed dosage not being exceeded, as explained above.

[0101] In the above passages, different ways of determining a portion-specific dosage have been explained.

[0102] In the following, generating control data for one or more operation modes will be described. The term operation mode is to be understood broadly and may refer to different modes of operation of an application device.

[0103] Specifically, the following operation modes differ in how the scenario is addressed where a threshold value for the sensor parameter is not exceeded by the sensor parameter value. This may be the case for one or more of the plurality of portions, for example, when coverage in a portion is low. Employing such a threshold value is optional. It may be advantageous for different reasons that will be explained below.

[0104] According to the present disclosure, the method may comprise generating the control data for one or more operation modes comprising a first operation mode and / or a second operation mode and / or a third operation mode.

[0105] For the first operation mode, e.g. a spiking mode, generating the control data may comprise: determining, for each of the plurality of portions, whether the sensor parameter value exceeds a predetermined threshold value, for each of the plurality of portions for which the sensor parameter value exceeds the predetermined threshold value, determining a variable portion-specific dosage depending on the sensor parameter value, and for each of the plurality of portions for which the sensor parameter value does not exceed the predetermined threshold value, setting the portion-specific dosage to zero.

[0106] In other words, in each portion, in a range of sensor parameter values above the predetermined threshold value, the dosage may have a sensor parameter value-dependency, i.e. , be a function of the sensor parameter (the function may, for example, include a step function), whereas in a range of sensor parameter values that do not exceed the predetermined threshold value, the dosage has no such dependency and, instead, in the present example, has a value of zero. The first operation mode results in no application of crop protection product in portions where the sensor parameter value does not exceed the threshold value.

[0107] The threshold value may, for example, be selected such that sensor parameter values below this threshold value are indicative of harmful organism characteristics that do not warrant applying crop protection product. For example, where the harmful organism is not of an aggressive, fast-spreading, and / or competitive type, below a certain coverage and / or abundance it may not be considered warranted to use crop protection product. The threshold value may then be selected to reflect said coverage and / or abundance. Different threshold values may be associated with respective harmful organisms and / or crop protection products. The threshold values may be retrieved from a data storage, for example based on the harmful organism present in the portion and / or the crop protection product.

[0108] Thus, in the first operation mode, use of crop protection product can be reduced without deteriorating crop protection.

[0109] In the first mode, when the threshold value is exceeded, a variable portion-specific dosage depending on the sensor parameter value is determined. This may be done in any of the ways described further above, for example based on an efficacy curve and a target efficacy associated with the sensor parameter value or based on relative dosage values associated with the sensor parameter value.

[0110] According to the present disclosure, the determination as to whether the sensor parameter value exceeds the predetermined threshold value and the respective setting of the dosage to zero or to the variable portion-specific parameter value is carried out in real-time. In this context, real-time may be considered to be the time when the application device moves through the agricultural field for applying the crop protection product. Particularly, the determination may take place in real time, such as when the respective portion of the agricultural field is sensed, i.e., sensor data are obtained for said portion.

[0111] As briefly mentioned above, the first mode may be suitable for certain harmful organisms or crop protection products. However, there may be harmful organisms, for example rather aggressive, fastspreading and / or competitive organisms, where it would be detrimental to the crop protection to operate in this first mode. There may also be crop protection products where it would be detrimental to the crop protection to operate in this first mode.

[0112] To address this aggressive, fast spreading and / or competitive type of harmful organism, the present disclosure may also provide a second operation mode.

[0113] For the second operation mode, e.g. a spotting mode, generating the control data may comprise: determining, for each of plurality of portions, whether the sensor parameter value exceeds a predetermined threshold value, for each of the plurality of portions for which the sensor parameter value exceeds the predetermined threshold value, determining a variable portion-specific dosage depending on the sensor parameter value, and for each of the plurality of portions for which the sensor parameter value does not exceed the predetermined threshold value, setting the portion-specific dosage to a predetermined baseline dosage value.

[0114] The second operation mode corresponds to the first operation mode except for how the scenario is handled that the threshold value is not exceeded, and reference is made to the above description.

[0115] Contrary to the first operation mode, the portion-specific dosage is set to a predetermined baseline dosage value, offset dosage value, or minimum dosage value when the sensor parameter value does not exceed the threshold. It is to be understood that the baseline value is non-zero, i.e. different from zero. In an example the baseline dosage value may substantially correspond to the dosage allocated to the threshold value.

[0116] Thus, rather than a variable portion-specific dosage, a fixed baseline dosage value is used when the threshold value is not exceeded. This allows, for example, for making sure that even for low coverage, dosage does not go below a minimum dosage. Thus, crops can be more reliably protected. Especially where organisms are fast spreading, this allows for reliable protection.

[0117] For the third operation mode, e.g. an automatic dosage detection mode, generating the control data comprises, for each of the plurality of portions of the field, determining a variable portion-specific dosage depending on the sensor parameter value. In this operation mode, a variable portion-specific dosage depending on the sensor parameter value is used, irrespective of any threshold value for the sensor parameter value.

[0118] Optionally, the operation mode may be set based on a user input selecting the operation mode. Alternatively, the operation mode may be automatically determined based on the crop protection product and / or one or more harmful organisms that are determined to be present or expected to be present in the agricultural field.

[0119] According to the present disclosure the method may comprise, when a plurality of types of harmful organisms are determined to be present in the portion or site of the agricultural field, selecting among the plurality of types the type having the highest dosage demand. The determining of the portion-specific dosage may be carried out under the assumption that substantially all of the harmful organisms in the respective portion are of the selected type. Thus, the risk under-dosing of the crop protection product is reduced. In particular, this may be the case where the harmful organisms are susceptible to the same tank mix.

[0120] The dosage demand of a type of harmful organism may be defined in any suitable manner. For example, the dosage demand may be a value derived from a data set comprising information on the dosage response for the respective type of harmful organism, such as a table or a dosage-response curve.

[0121] The determining whether a plurality of types of harmful organisms are present in the portion of the agricultural field may be based on sensor data, particularly from previously obtained sensor data and / or from real-time sensor data. Previously obtained sensor data may be sensor data that record a predefined time period before the current operation, e.g. current treatment of the field, for substantially the same portion of the field and / or site of the field. The determining that a plurality of types of harmful organisms are present in the portion of the agricultural field may alternatively or in addition be based on a user input indicating the types of harmful organisms.

[0122] As mentioned above, the determining whether a plurality of types of harmful organisms are present in the portion of the agricultural field may, in particular, be based on real-time sensor data. That is the determination may be made in real-time. This may be referred to as an online approach. Current sensor parameter and / or a current composition as well as a current fill level of a tank mix may be detected during the treatment on the treatment device in the field together with the received information from the server. The processing logic located on the treatment device may determine the dosage for the respective portion of the field and generates the dosage in form of control data for the treatment device, e.g. for a nozzle of the treatment device. Dose determination will be on the field in real time. If a plurality of harmful organisms are detected in a portion of the field, the harmful organism that is most dose demanding may be determined. In order to determine the most demanding harmful organism for each detected harmful organism a (general) dosage requirement may be determined and the one with the highest dosage demand may be selected and it may be assumed that all harmful organisms are of said type and treated cumulatively.

[0123] In contrast to the online approach, in an offline approach instead of detecting the type of the harmful organism by a sensor during the treatment of the field, the harmful organism may be provided by a user. Already before treatment the organism-specific dosage-response curve is provided and used for the whole field, independently from the characteristic of harmful organism that is detected by the sensor for determining the sensor parameter value.

[0124] Consequently, in the online approach different organism-specific dosage-response curves are selected dependent on the detected sensor parameter, in the offline approach a pre-selected dosage-response curve is used, in particular only one single organism-specific dosage-response curve is used for the whole field. In the online approach the organism-specific dosage-response curves may vary from portion to portion and / or from site to site dependent on the detected harmful organism.

[0125] In an example, the online approach may establish a communication connection which may be a connection to a server, e.g. via a wireless communication network and / or a cloud computing environment. Data necessary for processing, such as historical sensor data from a previous treatment campaign, may be downloaded as needed at substantially any time prior or during the treatment and as long as a network connection is available. Necessary information may be exchanged between treatment device and server during the treatment of the field in particular in an online approach. In an example, the organism-specific dosage-response curve, the decision logic and / or a dosage matrix for dose adaption may be sent to the treatment device.

[0126] The sensor parameter value for a portion of the agricultural field and the presence of a plurality of types of harmful organisms in the portion of the agricultural field may be determined based on the same sensor data, particularly from real-time sensor data.

[0127] The method of the present disclosure may entail defaulting to a user input indicating the type or types of harmful organisms in case an approach using the sensor data for determining whether a plurality of types of harmful organisms are present fails.

[0128] Determining the portion-specific dosage could be quite complicated if trying to distinguish and take into account different types of harmful organisms in each portion, particularly in real-time. The complexity of the determination can be decreased by assuming that all harmful organisms present in the portion are of the same type and determining the portion-specific dosage for this type of harmful organism.

[0129] Selecting, among the plurality of types, the type of harmful organism having the highest dosage demand, allows for reducing complexity as outlined above and at the same time providing reliable protection for the crop. Other selections of the type of harmful organism would run the risk of underdosing and providing unreliable crop protection. Accordingly, selecting the type of harmful organism having the highest dosage demand may be referred to as a conservative approach.

[0130] According to the present disclosure, the type of harmful organism may be derived from the sensor data, particularly from previously obtained sensor data and / or from real-time sensor data. In particular, the sensor parameter value for a portion of the agricultural field and the type of harmful organism in the portion of the agricultural field may be determined based on the same sensor data, particularly on real-time sensor data. Any sensor data processing techniques, for example image processing techniques, allowing for deriving the type of harmful organism from the sensor data as known in the art may be employed for this purpose.

[0131] Alternatively or in addition, the type of harmful organism may be determined on the basis of a user input indicating the type of harmful organism.

[0132] Determining the type of harmful organism from sensor data, for example image data, in real-time can most precisely reflect the current situation in the agricultural field. Where precision is required, such an approach is advantageous. However, such an approach is also computationally expensive. In some cases, the additional precision may not be required, or the available computational power may not be sufficient to carry out the computationally expensive approach in real-time. In such cases, the approach based on a user input and / or the approach based on previously obtained sensor-data is / are advantageous.

[0133] According to the present disclosure, the method may be configured for real-time detection of harmful organisms based on real-time sensor data and for real-time generating of control data. To that end, sensor data obtained for determining sensor parameter values may also be used as an input for a sensor data processing method that processes the sensor data so as to determine the presence and optionally type of harmful organism. Alternatively or in addition, additional sensor data may be obtained at the same time as the sensor data used for determining the sensor parameter values and may be used as an input for a sensor data processing method that processes the additional sensor data so as to determine the presence and optionally type of harmful organism. In particular, the data processing method that processes the sensor data so as to determine the presence and optionally type, of harmful organisms may entail image processing methods and may be carried out based on image data captured by a camera. The above allows for reliably determining harmful organisms present in the portion. Image processing methods may entail inputting image data into models, particularly machine learning models. As an example, a machine learning model may be trained on labelled images of a field or portions thereof, the labels indicating the presence and / or type of harmful organisms. Thus, in use, the machine learning model may process input image data and output a prediction on presence and / or type of harmful organisms. Other, e.g. non-machine learning models, are also conceivable.

[0134] Optionally, a single type of harmful organism is selected among all types of harmful organism detected for the portion of the agricultural field. The portion-specific dosage may be determined based on an assumption that all detected harmful organisms are of the selected type of harmful organism, particularly wherein, among a plurality of candidate types of harmful organism, a type having the highest dosage requirement is selected.

[0135] For example, it may first be determined which type of harmful organism in general has the highest dosage demand and then an assumption can be made that all of the harmful organisms in the respective portion or in the entire field are of said type. It is noted that the camera parameter value may be based on the overall amount of all harmful organisms, e.g. the harmful organism amount used for the determination is cumulative for the different types of harmful organisms.

[0136] This may allow for making the method less complex while still yielding reliable crop protection.

[0137] According to the present disclosure, the generating control data may comprise generating application control data for individually controlling each spray device of the application device to modify the spray device’s application rate so as to dispense an amount of crop protection product in accordance with the portion-specific dosage.

[0138] Alternatively or in addition, the generating control data may comprise converting the portion-specific dosage to obtain control values for duty cycles of each spray device of the application device, particularly the control values comprising a pulse width modulation value.

[0139] Alternatively or in addition, the control data may comprise generating control data for causing a selection of one or more spray devices to be activated among a plurality of spray devices of the application device.

[0140] According to the present disclosure, generating the control data is based on a current location of the application device. In other words, position data reflecting the current position of the application device may be used for generating the control data. This allows for determining the portion or portions where the spray device is located and, accordingly, allows for selecting the corresponding portion-specific dosage, particularly by taking into account relative position of the spray devices on the application device. For example, the control data for dose adaptation may be control data generated based on portion-specific dosage and a current location of the application device and relative position of the spray devices on the application device. Controlling the application device may comprise individually controlling the spray devices of the application device dependent on their location relative to the respective portion and its portion-specific dosage.

[0141] In view of the above, using the generated control data, the operation of the application device can be controlled to apply crop protection product with the determined portion-specific dosage.

[0142] According to the present disclosure, the sensor parameters may, for example, be indicative of at least one of morphological parameters of the harmful organism like height and / or width and / or leaf form and / or plant form and / or color characteristics, growth stage of the harmful organism, biomass of the harmful organism, coverage area of the harmful organism, and / or abundance of the harmful organism. Details in this regard have been described above.

[0143] As previously mentioned, according to the present disclosure, the portion-specific dosage may be determined based on a single one of the one or more harmful organisms. This significantly reduces complexity and, accordingly, computational resources required for determining the portion-specific dosage. Selecting the single one of the one or more harmful organisms has been described above in the context of selection among different types of harmful organisms. That is, for example, it is assumed that all harmful organisms are assumed to be of the same type, e.g. species, and their abundance or other characteristics reflected in the parameter value, may be added up. Thus, they will collectively be represented by a parameter value that reflects the added-up characteristics.

[0144] BRIEF DESCRIPTION OF THE DRAWINGS

[0145] In the following, the present disclosure is further described with reference to the enclosed figures:

[0146] Figs. 1a and 1 b illustrate an exemplary system in which the method of the present disclosure may be carried out.

[0147] Fig. 2 illustrates an exemplary method according to the present disclosure.

[0148] Figs. 3a to 3c illustrates examples of determining a portion-specific dosage according to the present disclosure.

[0149] Fig. 4 illustrates examples of generating control data according to the present disclosure.

[0150] DETAILED DESCRIPTION OF EMBODIMENTS

[0151] The following embodiments are mere examples for implementing the method, the system or application device disclosed herein and shall not be considered limiting.

[0152] In Figure 1a, an exemplary, not-to scale agricultural field 1 in a schematic top-view is illustrated. The agricultural field is sub-divided into portions 1a or sites 1a. In the present example, the portions are illustrated as rectangular in top-view. However, the agricultural field may be sub-divided in other ways, i.e. , other shapes and / or other number of portions. In this example, harmful organisms 2 are schematically indicated by hatched areas. Harmful organisms may for example be weeds.

[0153] A sensor 3 may be configured to provide sensor data. The sensor may, for example, be a camera. The camera may capture images of the agricultural field. In this case, the sensor data may be image data.

[0154] In Figure 1a, two sensors 3 are shown to illustrate different potential configurations of the sensor. One of the sensors is arranged at the application device 4 or treatment device 4. The other one is shown removed from the application device to indicate that it is independently movable, for example, mounted on a drone, or it is positioned at a fixed location as an in-field sensor.

[0155] According to the present disclosure, a sensor parameter value is derived for the respective portions of an agricultural field. This may be accomplished by processing the sensor data using known methods.

[0156] As an example, there are different indices derivable from sensor data known in the art. Such indices may be determined and used directly or after further processing to derive the sensor parameter value for the respective portion.

[0157] In the example of a camera, image data may be subject to image processing. The image processing may comprise detecting the harmful organisms in the image. Any suitable image processing methods known in the art may be employed. Such methods may entail shape recognition, color analysis, and / or machine learning methods.

[0158] As one example, a value for the coverage by the harmful organism may be determined for each portion 1a. As a non-limiting example, the camera parameter values may be “zero”, “low”, “medium” and “high” representing different amounts of coverage, e.g. representing different ranges of coverage, wherein the ranges are determined by thresholds.

[0159] Also schematically shown is an application device 4 having spray devices 4a, e.g. nozzles, arranged along a beam of the application device. The illustrated configuration is merely exemplary and other configurations of the application device are conceivable. For example, the number and arrangement of the spray devices may be different.

[0160] The application device may be configured to apply crop protection product via the spray devices. The application device is configured such that the amount of crop protection product applied by the application device is adjustable by controlling operation of the application device. Operation of the application device may be controlled by control data.

[0161] The control data, according to the present disclosure, comprises control data for dose adaptation for the application device.

[0162] Controlling the operation of the application device, particularly the spray devices, while the application device moves through the agricultural field allows for selectively applying different amounts of crop protection product to the different portions of the agricultural field.

[0163] To that end, the control data for dose adaptation may be control data generated based on portionspecific dosage and a current location of the application device and relative position of the spray devices on the application device. Thus, the operation of the application device may be controlled such that it operates to apply the portion-specific dosage for the portion at which it is currently located. Portion specific application may comprise individually controlling the spray devices.

[0164] As explained above, for example, the control data for dose adaptation may be control data generated based on portion-specific dosage and a current location of the application device and relative position of the spray devices on the application device. Portion-specific application may comprise individually controlling the spray devices of the application device dependent on their location relative to the spray device. On the respective portion, which may be one portion for each spray device, the portion-specific dosage will be applied.

[0165] The application device may have a first data processing system 4b, a memory 4c, and a communication interface 4d allowing for data exchange with one or more external data processing systems 5.

[0166] The method of the present disclosure may be entirely carried out by the first data processing system. Alternatively, a part of the method may be carried out by the one or more external data processing systems other parts of the method at the first data processing system. This is referred to as a distributed execution of the method. Data exchange between the first data processing system and the one or more external data processing systems, for example for a distributed execution, may be via the communication interface. The communication interface may allow for wireless data transfer, such as via a mobile communication network or a wireless LAN. Alternatively or in addition the communication interface may allow for wire-bound data transfer, such as via a cable or a USB drive. An exemplary wireless data communication is schematically indicated by the dashed line 6 in Figure 1a.

[0167] The present disclosure relates to a system 7 comprising a computing system configured to carry out the method of the present disclosure. The system may comprise the application device 4 and / or the sensor 3. The computing system may comprise the first processing system 4b and / or the one or more external processing systems 5. The computing system may optionally comprise the memory 4c and / or the communication interface. In case the computing system comprises the first and the one or more external processing systems, it may be referred to as a distributed computing system.

[0168] Figure 1 b illustrates in more detail potential data flows and interfaces between the application device 4 or treatment device 4 and the data processing system 5 (e.g. computing unit). Specifically, data 11 may be provided by the application device 4 or treatment device 4 to the processing system 5, where it is received by an input interface 8. Moreover, data 12 may be provided from the data processing system 5 to the application device 4 or treatment device 4 via, particularly via the output interface 10 of the data processing system 5. A processing unit 9 of the data processing system 5 may process data, e.g. received from the application or treatment device 4 and / or via a user input, and provide at least part of the processed data, e.g. as part of data 12, to the application or treatment device 4. This may be done using a wireless data connection or a wired data connection or using a non-volatile memory medium such as a USB stick.

[0169] Thus, for example, some operations may be carried out in a distributed manner either when the application device 4 or treatment device 4 moves through the field or in advance. For example, user input may be received at system 5 and data may then be processed, and part of the processed data may be provided to the application device 4 or treatment device 4.

[0170] In the following, methods according to the present disclosure will be described. These methods may, as an example, be carried out in a setting as described above. Particularly, a system as described above may be used to carry out these methods. Accordingly, reference will in some cases be made to the above system. However, it is to be understood that other systems may alternatively be used to carry out these methods.

[0171] A method for generating control data for dose adaptation for an application device for applying a crop protection product according to the present disclosure is schematically shown in Figure 2.

[0172] The method according to the present disclosure comprises, in Step S1 , obtaining, for each of a plurality of portions of an agricultural field, a sensor parameter value of a sensor parameter, wherein the sensor parameter value is derived from sensor data depicting the respective portion of the agricultural field and is indicative of harmful organism characteristics in the respective portion. Step S1 , for example, may be carried out during operation of the application device. More specifically, step S1 may be carried out while the application device moves through the agricultural field.

[0173] Sensor data may be obtained independently of the location of the application device relative to the respective portion in case the sensor data is obtained by a sensor movable independently of the application device. Sensor data will be obtained for portions that are within the current Field of View, FoV, of the sensor. In case the sensor is arranged at the application device, the Field of View depends on the location of application device and configuration and arrangement of the sensor.

[0174] The method according to the present disclosure comprises, in Step S2, determining, for each of the plurality of portions of the agricultural field, one or more harmful organisms present in the portion. This step may be carried out prior to, after, or during Step S1 .

[0175] As an example, in an offline approach Step S2 may entail receiving a user input specifying the one or more harmful organisms present in the portion prior to the application device starting operation. Alternatively, Step S2 may entail, prior to the application device starting operation, retrieving previously obtained data representative of one or more harmful organisms present in the portion. Alternatively, in an online approach Step S2 may be based on processing sensor data obtained during operation of the application device, for example the sensor data used for Step S1 . For example, the sensor data may be input into a model, as raw data or after preprocessing, and the harmful organism(s) may be determined by the model.

[0176] The method according to the present disclosure comprises, in Step S3, determining a portionspecific dosage for each of the plurality of portions on the sensor parameter value, on at least one of the one or more harmful organisms, and on the crop protection product.

[0177] Step S3 may be carried out during operation of the agricultural device, specifically, while the agricultural device moves through the agricultural field. Specifically, the portion-specific dosage may be determined in a real-time manner.

[0178] As mentioned above, the sensor data may be obtained by a sensor arranged at the application device. Sensor data will be obtained for portions that are within the current Field of View. The Field of View may substantially correspond to the detection region of a sensor. The Field of View of a camera may depend on the optic of the camera. The Field of View may correspond to a portion 1a of the field and / or comprise at least one portion 1a, a plurality of portions and / or parts of a portion 1a. In cases where the sensor is arranged on the application device, the Field of View may depend on the position of the application device. Accordingly, the portion-specific dosage may be determined upon the respective portion coming into the Field of View as the application device moves through the agricultural field. In a preferred embodiment, the Field of View depicts multiple portions, particularly all portions in the application range of the application device. The application device may comprise multiple sensors and sensor parameter values may be obtained from the multiple sensors. For example, where the Field of View does not depict all portions in the range of the application device, the application device may comprise multiple sensors whose combined Field of View depicts all portions in the range of the application device. In particular, there may be multiple sensors and among the multiple sensors one sensor or a subgroup of sensors may depict a single portion. That is, one portion may be handled by one or more associated sensors.

[0179] In case the image data is obtained by a sensor that is movable independently of the application device, such as a sensor mounted on a drone, the step of determining the portion-specific dosage may be decoupled from the movement of the application device. For example, this may allow for more processing time afforded for the processing of the sensor data. The method according to the present disclosure comprises, in Step S4, generating control data for dose adaptation for the application device based on the determined portion-specific dosage, and optionally a current location of the application device.

[0180] In particular, as explained above, the current location of the application device and the respective relative position of the spray devices on the application device may be taken into account when generating the control data. For example, the control data for dose adaptation may be control data generated based on portion-specific dosage and a current location of the application device and relative position of the spray devices on the application device. As an example, the control data may allow for individually controlling the spray devices of the application device dependent on their location, i.e. depending on which the corresponding portion is. The corresponding portion is the portion in which the location of the spray device is. One or multiple portions may be present per application device. On the corresponding portion the portion specific dosage will be applied.

[0181] The control data is designated for being used to control the application of the crop protection product. The method may optionally comprise Step S5 of controlling operation of the application device based on the control data.

[0182] In the following different ways of carrying out Step S3 will be explained, i.e., different methods of determining the portion-specific dosage, making reference to Figures 3a to 3c. These methods can also be combined. The different methods are also referred to as different determination modes for determining the portion-specific dosage, particularly where a selection between the different methods is provided. Such a selection may be made automatically or via a user input.

[0183] The method may optionally comprise, in Step S31 , a selection between the determination modes. Alternatively, the method may comprise only the first or only the second determination mode.

[0184] The first method or determination mode for determining the portion-specific dosage is based on a predetermined target efficacy. This first determination mode may be referred to as “site specific spotting”. The predetermined target efficacy may be sensor-parameter-value-specific. This means that, for different sensor parameter values, there is a respective associated target efficacy. For example, a database may provide information specifying, for different sensor parameter values (or ranges thereof), the respective associated target efficacy.

[0185] The sensor parameter values, which may be raw values from sensors or derived parameter values, may reflect or be an indication of a harmful organism growth state, e.g. a coverage area. Depending on the growth state, the required dosage for successfully treating the field, e.g. to eradicate the harmful organism, will differ. Determining a suitable dosage may employ target efficacies.

[0186] For example, from the detected growth state, reflected by the sensor parameter values, a target efficacy (desired efficacy) for the detected growth state may be derived, such as retrieved from a database. The target efficacies may be range-based target efficacies, e.g. associated with a respective range of growth states. The ranges may be associated, respectively, with sensor parameter values “zero”, “low”, “medium”, or “high”, for example. As explained above, the target efficacy values each have a corresponding dosage, derivable, for example, for the efficacy curve. A high target efficacy may have a corresponding high dosage, for example.

[0187] Determination of a dosage may, as an example, be based on the efficacy curve, which provides efficacy as a function of dosage. As an example, in order to determine, for different growth states, a dosage, a respective target efficacy may be assigned to different growth states, for example to different sensor parameter values or sensor parameter value ranges. The target efficacy has a corresponding dosage in the efficacy curve. Accordingly, via the dosage efficacy curve and a target efficacy, for any given sensor parameter value determined during operation, a corresponding target efficacy can be determined and the dosage corresponding to this target efficacy can be determined. The first method comprises selecting a dosage value corresponding to the target efficacy based on an organism-specific dosage-response curve. Each organism-specific dosage-response curve is associated with a respective harmful organism. The organism-specific dosage-response curve may be retrieved from a memory device.

[0188] The organism-specific dosage-response curve represents the efficacy of the crop protection product for a given type of harmful organism as a function of dosage. For example, where the harmful organisms are weeds, each weed may have an associated dosage-response curve.

[0189] In more detail, the portion specific dosage according to the first method or determination mode for determining the portion-specific dosage comprises, in Step S32, retrieving the predetermined target efficacy associated with the sensor parameter value, in particular associated with a sensor parameter related to a growth state of the harmful organism. The method further comprises, in Step S33, retrieving the organism-specific dosage-response curve representing the efficacy of the crop protection product for a given type of harmful organism. The method further comprises, in Step S34, determining the dosage value associated with the target efficacy from the dosage-response curve. The method further comprises, in Step S35, selecting said dosage value as portion-specific dosage.

[0190] Where only one harmful organism is determined to be present in the portion, this will yield exactly one portion-specific dosage. Where different harmful organisms are determined to be present in the portion, different portion-specific dosages may be determined from the different dosage-response curves. This scenario will be explained in more detail further below.

[0191] An exemplary dosage-response curve is shown in Figure 3b. In this example, the dosage-response curve is the dosage-response curve for the weed ALOMY. Vertical lines A, B, C, and D represent four sensor parameter values, e.g. values derived by processing raw data from a sensor, such as a camera. These sensor parameter values are, merely as an example, referred to as “zero”, “low”, “medium”, and “high”. The sensor parameter value may represent four categories of growth state and / or extent of coverage of the weed. Each sensor parameter value has an associated target efficacy, Teff, which can be derived from the dosage-response curve. This is illustrated in Figure 3b by the vertical lines intersecting with the dosage-response curve at the respective target efficacy and corresponding dosage value.

[0192] An exemplary table illustrating dosages for this example, resulting from the first method or determination mode for determining the portion-specific dosage, is inserted below:

[0193] In other words, the dosage-response curve of Figure 3b shows a relation between dosage values in l / ha on the abscissa and efficacy values from 0 to 100%. The dosage-response curve is a S-shaped curve. Furthermore, Figure 3b shows borders of ranges and / or threshold values A, B, C, D. The threshold value The threshold value A indicates a dosage value 1 .45 l / ha and efficacy value of 30%, The threshold value B indicates a dosage value 1.7 l / ha and efficacy value of 75%, The threshold value C indicates a dosage value 1 .8 l / ha and efficacy value of 85%, The threshold value D indicates a dosage value 2.2 l / ha and efficacy value of 95%, In this way ranges of the efficiency values can be mapped to corresponding dosage values. The ranges may be associated with the sensor parameter values “zero”, “low”, “medium”, “high”. If the sensor parameter value is zero (e.g. optionally, this may involve that a value is withing a range of values and the range is defined as corresponding to sensor parameter value “zero”), an associated target efficacy based on the above table, is 30%. This similarly applies for the other sensor parameter values low, medium, and high, and their associated target efficacy (75 %, 85 %, and 95%, respectively, in the above table). Each target efficacy, via one of the values A, B, C, D is associated with a corresponding dosage. For example, if, for a sensor parameter value zero it is desired to have an efficacy of 30% a dosage of 1 .45 l / ha is to be selected. If, for a sensor parameter value low it is desired to have an efficacy of 75%, a dosage of 1.7 l / ha is to be selected. If, for a sensor parameter value medium it is desired to have an efficacy of 85% a dosage of 1 .8 l / ha is to be selected, and if, for a sensor parameter high, it is desired to have an efficacy of 95%, a dosage of 2.2 l / ha is to be selected.

[0194] It is noted that in this example values are grouped together into sensor parameter values Zero, Low, Medium, and High, and respective target efficacies are provided. However, it is similarly possible that sensor parameter values are not such groups of values. That is, sensor parameter values may be a continuous range of values. In this case, for each of the continuous range of values (sensor parameter values), there may be an associated target efficacy.

[0195] The second method or determination mode for determining the portion-specific dosage comprises, in step S36, retrieving a predetermined relative dosage value, such as a percentage, associated with the sensor parameter value. The method further comprises, in step S37, determining the portionspecific dosage by multiplying the relative dosage value with a baseline dosage value.

[0196] The relative dosage value and / or the baseline dosage value may optionally be harmful-organismspecific. Thus, where different harmful organisms are determined to be present in the portion, multiple dosage values may thus be determined. This scenario will be explained further below.

[0197] An exemplary table illustrating dosages resulting from the second method or determination mode for determining the portion-specific dosage is inserted below:

[0198] Data as represented by the tables above may be obtained prior to operation of the application device and stored in memory device of the application device for retrieval, for example as a matrix. As a part of generating the control data, the sensor parameter value may be determined, and a corresponding dosage may be determined by retrieving the data stored in the memory device and selecting the dosage value associated with the sensor parameter value.

[0199] Figure 3c illustrates optional variations of the method of Figure 3a. The method, in this case, may comprise determining, whether a single or a plurality of types of harmful organisms is present in the portion of the agricultural field.

[0200] When a plurality of types of harmful organisms is determined to be present in the portion of the agricultural field, the method may comprise selecting, in Step S30, among the plurality of types, the type having the highest dosage demand. The determining of the portion-specific dosage is carried out under the assumption that all of the harmful organisms in the respective portion is of the selected type. For example, where multiple types of different harmful organisms are present in a portion, multiple respective candidate dosage values may be determined for a single sensor parameter value in Steps S31 to S37.

[0201] The method may comprise, determining in, Step S30, the harmful organism having the highest dosage demand among the multiple different harmful organisms, and determining, in Steps S31 to S37, a portion-specific dosage only for the determined harmful organism. It is noted that the sensor parameter value may be based on the overall amount of all harmful organisms, e.g. the harmful organism amount used for the determination is cumulative for the different harmful organisms. In other words, it may first be determined which pest (organism type) in general has the highest demand and then an assumption is made that all of the harmful organisms in the respective portion are of said type, i.e. it is then the selected type.

[0202] Step S30 outlined above provides the most reliable protection of the crop. Alternatively, other predetermined rules may be applied to arrive at a single portion-specific dosage where multiple different types of harmful organisms are present.

[0203] Where the above-described different modes are provided, optionally the mode may be automatically selected based on the type of harmful organism.

[0204] Figure 4 illustrates method steps for generating control data, for example as part of Step S4 described above. Control data may be generated for one or more operation modes. The method of the present disclosure may comprise, in Step S41 selecting an operation mode prior to generating the control data, automatically or via a user input.

[0205] As an example, generating the control data may comprise generating control data for a first operation mode. The method comprises, in Step S42, determining, for each of the plurality of portions, whether the sensor parameter value exceeds a predetermined threshold value. The method further comprises, in Step S43, for each of the plurality of portions for which the sensor parameter value exceeds the predetermined threshold value, determining a variable portion-specific dosage depending on the sensor parameter value. The method comprises, in Step S44, for each of the plurality of portions for which the sensor parameter value does not exceed the predetermined threshold value, setting the portion-specific dosage to zero. Accordingly, no crop protection product would be applied where the sensor parameter value does not exceed the threshold value. Otherwise, a variable portion-specific dosage is determined. The variable portion-specific dosage may be based on a dosage-response curve or based on a relative dosage value, for example, as outlined above in the context of Figures 3a to 3c.

[0206] For example, continuing the above example, the sensor parameter value “Zero” may represent a predetermined threshold value for the sensor parameter value. For example, where a sensor parameter is a parameter representative of an amount of harmful organism present in the portion (e.g. determined by image processing), sensor parameter value “Zero” may indicate that there is no harmful organism present and / or that its amount remains below a certain threshold. The values “Low”, “Medium”, and “High” exceed this value. Control data may be generated such that for any sensor parameter value that does not exceed the threshold value, the dosage to be applied will be set to a fixed value, in particular, to zero. Methods of applying crop protection product wherein the fixed value is set to zero are also referred to as “Spotting”. In such a case, continuing the above example, the table for the dosage to be applied would be as follows:

[0207] As another example, generating the control data may comprise generating control data for a second operation mode. The method comprises, in Step S42, determining, for each of the plurality of portions, whether the sensor parameter value exceeds a predetermined threshold value. The method also comprises, in Step S45, for each of the plurality of portions for which the sensor parameter value exceeds the predetermined threshold value, determining a variable portion-specific dosage depending on the sensor parameter value. This step corresponds to Step S43 described above. The method comprises, in Step S46, for each of the plurality of portions for which the sensor parameter value does not exceed the predetermined threshold value, setting the portion-specific dosage to a predetermined baseline dosage value. The baseline dosage value may be predefined, for example retrieved from a memory device or obtained by a user input. The baseline dosage value in this context is to be understood as being non-zero.

[0208] As another example, generating the control data may comprise generating control data for a third operation mode. The method comprises, in Step S47, for each of the plurality of portions, determining a variable portion-specific dosage depending on the sensor parameter value. This mode is a mode wherein the dosage entirely independent on whether or not a threshold value is exceeded.

[0209] The threshold may be calculated on field meta data and is based on field trials. It tries to optimize for an acceptable weed control of the farmer, in other words efficacy. For example, the threshold may be set to coverage low and only once the sensor parameter low is exceeded a crop protection product is applied. Known methods for determining the threshold may be used.

[0210] In the above examples, optionally, a maximum allowed dosage to be applied may be taken into account. For example, such a maximum allowed dosage may be based on regulatory limitations. Where said maximum allowed dosage would be exceeded by the dosage value from the dosageresponse curve, the dosage from the dosage-response curve may be replaced with the maximum allowed dosage as a dosage to be applied. Continuing the example described above and using an exemplary maximum allowed dosage of 2.1 l / ha, the table for the dosage to be applied would be as follows:

[0211] Further examples and embodiments will be outlined below making reference to the above descriptions.

[0212] The method of the present disclosure allows for avoiding applying excess crop protection product while still reliably achieving control of harmful organisms. Moreover, crop protection product will be better distributed over the agricultural field. Particularly, dosage may be reduced in some parts of the agricultural field without impeding crop protection in other parts of the agricultural field, where the required dosage may be higher.

[0213] The method of the present disclosure allows for using one single sensor parameter value for a portion of an agricultural field to determine a portion-specific dosage reliably. This sensor parameter value can be determined in real-time from sensor data obtained by a sensor, e.g. a sensor mounted on the application device or an individually moving sensor. Further required is an identification of a harmful organism that is (expected to be) present in the portion. This may also be determined in real-time from the sensor data, or it may be based on a preceding input by a user. Among multiple input or detected types of harmful organisms, a single type may be selected. For example, the type requiring the highest dosage may be selected. Further required is information on the crop protection product being used.

[0214] As such, a combination of three parameters, only one of which strictly requires processing of live sensor data, is sufficient for determination of a portion-specific dosage.

[0215] An application device may be provided with a data set linking, directly or indirectly, multiple sensor parameter values with a respective dosage for a specific crop protection product and for a respective harmful organism. As outlined above, the data set associating each sensor parameter value with a respective target efficacy and retrieving a dosage from a dosage-response curve describing the response of a specific harmful organism to a crop protection product may be used. Alternatively or in addition, the data set associating each sensor value with a respective relative dosage value to be multiplied with a (non-zero) baseline dosage value that may depend on the harmful organism and crop protection product may be used.

[0216] Thus, for example, based only on a single the sensor parameter, a single harmful organism present, and the crop protection product, a portion-specific dosage for a portion of the agricultural field can be determined.

[0217] Since only the sensor parameter value needs to be determined in real-time and all other required information may be generated prior to operation of the application device, there is a greater flexibility in terms of computing resources required to allow for the real-time generation of control data.

[0218] As an example, the corresponding dosage for a given sensor parameter value may be determined and entered in a matrix. Agronomic expertise is the basis for the content of the matrix. The relationship of the crop protection product, the harmful organism, and the sensor parameter, and dosage may be entered into the matrix. The determination of the harmful organisms, for example species detection, may comprise using models to detect the harmful organism from the sensor data, which is referred to as online approach. Alternatively, the harmful organism can be entered by the user, which is referred to as offline approach. Where multiple harmful organisms are input or detected, the most conservative dosage may be selected.

[0219] Examples

[0220] 1 . A computer-implemented method for generating control data for dose adaptation for an application device (4) for applying a crop protection product, the method comprising obtaining, for each of a plurality of portions (1 a) of an agricultural field (1 ), a sensor parameter value of a sensor parameter, wherein the sensor parameter value is derived from sensor data depicting the respective portion (1a) and is indicative of harmful organism characteristics in the respective portion (1a); determining, for each of the plurality of portions (1a), one or more harmful organisms present in the portion (1a); determining a portion-specific dosage for each of the plurality of portions (1a) based on the sensor parameter value, on at least one of the one or more harmful organisms present in the portion, and on the crop protection product; and generating control data for dose adaptation for the application device (4) based on the determined portion-specific dosage.

[0221] 2. The method of example 1 , wherein the portion-specific dosage is determined based on a predetermined target efficacy, and comprises selecting a dosage value corresponding to the target efficacy based on an organism-specific dosage-response curve representing the efficacy of the crop protection product for a given type of harmful organism, growth stage of the harmful organism, biomass of the harmful organism, coverage area of the harmful organism, and / or abundance of the harmful organism.

[0222] 3. The method of example 1 or 2, wherein the method comprises one or more determination modes for determining the portion-specific dosage comprising a first determination mode and / or a second determination mode, wherein determining the portion-specific dosage using the first determination mode comprises: retrieving a / the predetermined target efficacy associated with the sensor parameter value, retrieving a / the organism-specific dosage-response curve representing the efficacy of the crop protection product for a given type of harmful organism, determining the dosage value associated with the target efficacy from the dosage-response curve, and selecting said dosage value as portion-specific dosage, and / or wherein determining the portion-specific dosage using the second determination mode comprises: retrieving a predetermined relative dosage value associated with the sensor parameter value or absolute dosage value associated with the sensor parameter value, and if a relative dosage value is retrieved, the second determination mode may further comprise determining the portion-specific dosage by multiplying the relative dosage value with a baseline dosage value.

[0223] 4. The method of any of the preceding examples, wherein the method comprises generating the control data for one or more operation modes comprising a first operation mode and / or a second operation mode and / or a third operation mode, wherein, for the first operation mode, generating the control data comprises: determining, for each of the plurality of portions (1a), whether the sensor parameter value exceeds a predetermined threshold value, for each of the plurality of portions (1a) for which the sensor parameter value exceeds the predetermined threshold value, determining a variable portion-specific dosage depending on the sensor parameter value, and for each of the plurality of portions (1a) for which the sensor parameter value does not exceed the predetermined threshold value, setting the portion-specific dosage to zero, and / or wherein, for the second operation mode, generating the control data comprises: determining, for each of plurality of portions, whether the sensor parameter value exceeds a predetermined threshold value, for each of the plurality of portions (1a) for which the sensor parameter value exceeds the predetermined threshold value, determining a variable portion-specific dosage depending on the sensor parameter value, and for each of the plurality of portions (1a) for which the sensor parameter value does not exceed the predetermined threshold value, setting the portion-specific dosage to a predetermined baseline dosage value, and / or wherein, for the third operation mode, generating the control data comprises, for each of the plurality of portions (1a), determining a variable portion-specific dosage depending on the sensor parameter value.

[0224] 5. The method of any of the preceding examples, wherein the method comprises, when a plurality of types of harmful organisms is determined to be present in the portion (1a), selecting, among the plurality of types, the type having the highest dosage demand, and wherein the determining of the portion-specific dosage is carried out under the assumption that all of the harmful organisms in the respective portion (1a) is of the selected type.

[0225] 6. The method of any of the preceding examples, wherein the type of harmful organism is derived from sensor data, particularly from previously obtained sensor data and / or from real-time sensor data, particularly the sensor data from which the sensor parameter value is derived, and / or is determined on the basis of a user input indicating the type, particularly the species, of harmful organism.

[0226] 7. The method of any of the preceding examples, wherein the method is configured for real-time detection of harmful organisms based on real-time sensor data and real-time generating of control data, particularly wherein a single type of harmful organism is selected among all types of harmful organisms detected for the portion (1a), and / or wherein the portion-specific dosage is determined based on an assumption that all detected harmful organisms are of the selected type of harmful organism, particularly wherein among a plurality of candidate types of harmful organisms a type having the highest dosage requirement is selected.

[0227] 8. The method of any of the preceding examples, wherein generating the control data comprises generating application control data for individually controlling each spray device (4a) of the application device (4) to modify the spray device’s application rate so as to dispense an amount of crop protection product in accordance with the portion-specific dosage, and / or converting the portion-specific dosage to obtain control values for duty cycles of each spray device (4a) of the application device, particularly the control values comprising a pulse width modulation value, and / or generating application control data for causing a selection of one or more spray devices (4a) to be activated among a plurality of spray devices (4a) of the application device (4).

[0228] 9. The method of any of the preceding examples, wherein generating the control data is based on a current location of the application device (4) and optionally a relative position of the application devices on the application device.

[0229] 10. The method of any of the preceding examples, wherein the sensor parameter comprises or is derived from at least one of: morphological parameters of the harmful organism like height and / or width and / or leaf form and / or plant form and / or color characteristics, growth stage of the harmful organism, biomass of the harmful organism, coverage area of the harmful organism.

[0230] 11 . The method of any of the preceding examples, wherein the portion-specific dosage is determined based on a single one of the one or more harmful organisms.

[0231] 12. Use of the control data generated by the method of any one of examples 1 to 11 for controlling an application device (4) to apply crop protection product in accordance with the control data. 13. A system (7) comprising a computing system (4b, 5) configured to carry out the method of any one of examples 1 to 11 .

[0232] 14. A computer program product comprising instructions which, when carried out by a computer, cause the computer to carry out and / or control the method of any one of examples 1 to 11 .

[0233] 15. A computer-readable medium having stored thereon instructions which, when carried out by a computer, cause the computer to carry out and / or control the method of any one of examples 1 to 11.

[0234] The present disclosure has been described in conjunction with preferred embodiments and examples as well. However, other variations can be understood and effected by those persons skilled in the art and practicing the claimed invention, from the studies of the drawings, this disclosure, and the claims. Notably, unless otherwise specified, the any steps presented can be performed in any order, i.e., the present invention is not limited to a specific order of these steps. Moreover, unless specified otherwise, it is also not required that the different steps are performed at a certain place or at one computing device of a distributed system, i.e., each of the steps may be performed at different computing devices.

Claims

CLAIMS1 . A computer-implemented method for generating control data for dose adaptation for an application device (4) for applying a crop protection product, the method comprising obtaining, for each of a plurality of portions (1 a) of an agricultural field (1 ), a sensor parameter value of a sensor parameter, wherein the sensor parameter value is derived from sensor data depicting the respective portion (1a) and is indicative of harmful organism characteristics in the respective portion (1a); determining, for each of the plurality of portions (1a), one or more harmful organisms present in the portion (1a); determining a portion-specific dosage for each of the plurality of portions (1a) based on the sensor parameter value, on at least one of the one or more harmful organisms present in the portion, and on the crop protection product; and generating control data for dose adaptation for the application device (4) based on the determined portion-specific dosage, wherein the portion-specific dosage is determined based on a predetermined target efficacy, and the method comprises selecting a dosage value corresponding to the target efficacy based on an organism-specific dosage-response curve.

2. The method of claim 1 , wherein the organism-specific dosage-response curve is representing the efficacy of the crop protection product for a given type of harmful organism, growth stage of the harmful organism, biomass of the harmful organism, coverage area of the harmful organism, and / or abundance of the harmful organism.

3. The method of claim 1 or 2, wherein the method comprises one or more determination modes for determining the portion-specific dosage comprising a first determination mode and / or a second determination mode, wherein determining the portion-specific dosage using the first determination mode comprises: retrieving a / the predetermined target efficacy associated with the sensor parameter value, retrieving the organism-specific dosage-response curve, the dosage-response curve particularly representing the efficacy of the crop protection product for a given type of harmful organism, determining the dosage value associated with the target efficacy from the dosage-response curve, and selecting said dosage value as portion-specific dosage, and / or wherein determining the portion-specific dosage using the second determination mode comprises: retrieving a predetermined relative dosage value associated with the sensor parameter value or absolute dosage value associated with the sensor parameter value, and if a relative dosage value is retrieved, the second determination mode may further comprise determining the portion-specific dosage by multiplying the relative dosage value with a baseline dosage value.

4. The method of any of the preceding claims, wherein the method comprises generating the control data for one or more operation modes comprising a first operation mode and / or a second operation mode and / or a third operation mode,wherein, for the first operation mode, generating the control data comprises: determining, for each of the plurality of portions (1a), whether the sensor parameter value exceeds a predetermined threshold value, for each of the plurality of portions (1a) for which the sensor parameter value exceeds the predetermined threshold value, determining a variable portion-specific dosage depending on the sensor parameter value, and for each of the plurality of portions (1a) for which the sensor parameter value does not exceed the predetermined threshold value, setting the portion-specific dosage to zero, and / or wherein, for the second operation mode, generating the control data comprises: determining, for each of plurality of portions, whether the sensor parameter value exceeds a predetermined threshold value, for each of the plurality of portions (1a) for which the sensor parameter value exceeds the predetermined threshold value, determining a variable portion-specific dosage depending on the sensor parameter value, and for each of the plurality of portions (1a) for which the sensor parameter value does not exceed the predetermined threshold value, setting the portion-specific dosage to a predetermined baseline dosage value, and / or wherein, for the third operation mode, generating the control data comprises, for each of the plurality of portions (1a), determining a variable portion-specific dosage depending on the sensor parameter value.

5. The method of any of the preceding claims, wherein the method comprises, when a plurality of types of harmful organisms is determined to be present in the portion (1a), selecting, among the plurality of types, the type having the highest dosage demand, and wherein the determining of the portion-specific dosage is carried out under the assumption that all of the harmful organisms in the respective portion (1a) is of the selected type.

6. The method of any of the preceding claims, wherein the type of harmful organism is derived from sensor data, particularly from previously obtained sensor data and / or from real-time sensor data, particularly the sensor data from which the sensor parameter value is derived, and / or is determined on the basis of a user input indicating the type, particularly the species, of harmful organism.

7. The method of any of the preceding claims, wherein the method is configured for real-time detection of harmful organisms based on real-time sensor data and real-time generating of control data, particularly wherein a single type of harmful organism is selected among all types of harmful organisms detected for the portion (1a), and / or wherein the portion-specific dosage is determined based on an assumption that all detected harmful organisms are of the selected type of harmful organism, particularly wherein among a plurality of candidate types of harmful organisms a type having the highest dosage requirement is selected.

8. The method of any of the preceding claims, wherein generating the control data comprises generating application control data for individually controlling each spray device (4a) of the application device (4) to modify the spray device’s application rate so as to dispense an amount of crop protection product in accordance with the portion-specific dosage, and / orconverting the portion-specific dosage to obtain control values for duty cycles of each spray device (4a) of the application device, particularly the control values comprising a pulse width modulation value, and / or generating application control data for causing a selection of one or more spray devices (4a) to be activated among a plurality of spray devices (4a) of the application device (4).

9. The method of any of the preceding claims, wherein generating the control data is based on a current location of the application device (4) and optionally a relative position of the application devices on the application device.

10. The method of any of the preceding claims, wherein the sensor parameter comprises or is derived from at least one of: morphological parameters of the harmful organism like height and / or width and / or leaf form and / or plant form and / or color characteristics, growth stage of the harmful organism, biomass of the harmful organism, coverage area of the harmful organism.11 . The method of any of the preceding claims, wherein the portion-specific dosage is determined based on a single one of the one or more harmful organisms.

12. Use of the control data generated by the method of any one of claims 1 to 11 for controlling an application device (4) to apply crop protection product in accordance with the control data.

13. A system (7) comprising a computing system (4b, 5) configured to carry out the method of any one of claims 1 to 11 .

14. A computer program product comprising instructions which, when carried out by a computer, cause the computer to carry out and / or control the method of any one of claims 1 to 11 .

15. A computer-readable medium having stored thereon instructions which, when carried out by a computer, cause the computer to carry out and / or control the method of any one of claims 1 to 11 .

Images